Structured data is standardized markup, most often JSON-LD using the schema.org vocabulary, that describes a page’s content to machines: who you are, what you sell, who wrote the article. It guarantees neither a rich result nor a citation by an AI, but it removes ambiguity about who you are.
What structured data is
On your home page, a person instantly understands that the name at the top is your company’s, that the address in the footer is yours and that the phone number is how to reach you. A machine guesses. Structured data replaces that guesswork with labels.
Google defines it as “a standardized format for providing information about a page and classifying the page content.” The most widely used vocabulary is schema.org, founded by Google, Microsoft, Yahoo and Yandex. According to schema.org, more than 45 million web domains were using it in 2024.
There are three formats: JSON-LD, Microdata and RDFa. For Google, all three are equally fine as long as the markup is valid. Google still recommends JSON-LD, the easiest solution to implement and maintain at scale (Google Search Central). It is the format we recommend: one block of code in the page, separate from the text, that you can fix without touching the layout.
Examples: the schema types that matter for a small or mid-sized business
There is no need to mark up everything schema.org allows. For a small or mid-sized business, six types do most of the work:
- Organization or LocalBusiness: your name, logo, contact details and official profiles. It is your brand’s ID card.
- Article, with an author of type Person: who wrote what, and when. It is the foundation of a piece of content’s credibility.
- Service or Product: what you sell, described with the same words everywhere.
- BreadcrumbList: where the page sits in your site.
- FAQPage: your questions and answers. More on this below, because Google changed the rules in 2026.
- WebSite: your site’s name and language.
Here is what a minimal Organization block looks like. The values are fictitious: replace them with your own.
<script type="application/ld+json">
{
"@context": "https://schema.org",
"@type": "Organization",
"name": "Your Company Inc.",
"url": "https://www.example.ca",
"logo": "https://www.example.ca/logo.png",
"telephone": "+1-514-555-0100",
"address": {
"@type": "PostalAddress",
"addressLocality": "Montreal",
"addressRegion": "QC",
"addressCountry": "CA"
},
"sameAs": [
"https://www.linkedin.com/company/your-company"
]
}
</script>The most underrated line is the last one. The sameAs property links your site to your other official presences. It tells the machine: “the company on this site and the one on this LinkedIn page are the same.”
What changed in 2026: the end of FAQ rich results
For years, FAQPage markup was sold for one reason: getting expandable questions under your link in Google. That reason is gone.
Google had been limiting FAQ rich results since 2023. It stopped showing them on May 7, 2026 (Search Engine Journal, May 10, 2026), then removed their documentation on June 15, 2026 (Google Search Central changelog). The markup can stay on your pages without causing problems: it simply no longer produces any special display in Google.
Should you drop FAQs altogether? No. A question asked the way your customers ask it, followed by a short, self-contained answer, remains a format that answer engines extract easily. It is a criterion in its own right in our grid, the question-and-answer format, which carries 15% of the score. What changes is the promise: FAQPage markup no longer buys you space on the results page.
What about AI engines?
That is the real question, and the answer calls for rigour.
- Google states in black and white that no special schema.org structured data is needed to appear in AI Overviews or AI Mode. It only recommends that your structured data match the visible text on the page (“AI features and your website,” updated December 10, 2025).
- Microsoft said the opposite, or nearly. In March 2025, at SMX Munich, Fabrice Canel of the Bing team said that schema markup helps Microsoft’s large language models understand your content (Search Engine Land, March 20, 2025). His remarks covered Bing and Copilot.
- OpenAI and Perplexity: we found no public statement on their use of structured data. So we will not tell you it counts, or that it does not.
So why does structured data carry 20% of our GEO score (also called AEO), more than any other criterion? Because our score does not measure a shortcut. It measures what your site offers machines. And structured data is the only signal you fully control that states, without ambiguity, who you are, what you do and who is speaking. An engine torn between two companies with the same name, or between your price and a reseller’s, is less inclined to cite you. Our eight criteria and their weights are published. Learn more: our GEO score.
Test your structured data in 10 minutes
- Google’s Rich Results Test (search.google.com/test/rich-results): paste a page’s address. It tells you which markup Google can use for its displays, and which contains errors.
- The schema.org validator (validator.schema.org): it reads every type, including those Google does not use. This is where you see your Organization block or your author.
- Google Search Console: the enhancement reports track errors over time, across the whole site.
- The consistency check: your name, address and phone number must be identical in the markup, in the page text and on your official profiles. Two versions of your address are two companies to a machine.
- The honesty check: Google’s rule is simple, “don’t mark up content that is not visible to readers of the page” (general guidelines, updated July 10, 2026). Customer reviews or a star rating that do not appear on the page have no place in the code.
On WordPress, a plugin like Rank Math generates much of this markup. Check that it does not generate it a second time alongside your theme: two contradictory Organization blocks do more harm than none. Learn more: our SEO approach.
How we see it
Structured data is not a trick. Google says so itself: it does not guarantee that your structured data will show up in search results, even when correctly marked up. Its own case studies show what it can bring when it produces a rich display: a 25% higher click-through rate on Rotten Tomatoes pages with structured data, and an 82% higher click-through rate on Nestlé pages shown as rich results. These are click measurements in classic results, published by Google, not measurements of AI citations.
Our position: mark up what is true, visible and consistent from one page to the next, and stop there. Honest markup ages well, whichever engine reads it. That is how we see things, and we are not going to change. Learn more: our approach to AI search optimization.
Clear. Consistent. Verifiable.
Frequently asked questions
What is an example of structured data?
The most common one is a JSON-LD block of type Organization, placed in the code of your home page. It states your company’s name, address, logo, phone number and official profiles, such as LinkedIn. Other types describe an article and its author, a product and its price, a service, or a page’s breadcrumb trail.
Is structured data required to appear in Google's AI Overviews?
No. Google states that no special schema.org markup is required for AI Overviews or AI Mode, and recommends that your structured data match the visible text. Microsoft, for its part, says schema markup helps its language models understand your content. Markup is not an entry ticket: it is a way to remove ambiguity.
Should I remove FAQPage markup now that Google no longer shows FAQ rich results?
Nothing requires you to. Google stopped showing FAQ rich results in May 2026, and the markup can stay without causing problems. Above all, keep your questions and answers visible on the page: that format, written the way your customers ask, is what answer engines extract easily.
How do I test my structured data?
Use Google’s Rich Results Test to see what Google can use, and the schema.org validator to check every type, including those Google does not use. Then track errors in Google Search Console’s enhancement reports, and check that the markup describes what the page actually shows.
JSON-LD, Microdata or RDFa: which should I choose?
For Google, all three formats are equally fine as long as the markup is valid. Google still recommends JSON-LD, because it is the easiest to implement and maintain at scale. It is also the format we recommend: the code block stays separate from the text, and you can fix it without touching the layout.
